The Heart of Artificial Intellegence

    Ant colonies are smarter than you. For what they’ve got to work with, they solve some pretty intense problems and they solve them with style. They build lavish homes for themselves with designs that accommodate the different jobs they delegate out to each other. They’ve got builders, scavengers, warriors, queens and if there gets to be too many ants doing one kind of job, they change who does what until they’re back in balance again. It seems safe to say that colonies of people are nowhere near as smoothly successful, and that’s before we even start talking about how much of this stuff your dumb ass can do. Ants: one. You: zilch. 

Git ‘er done. 

     But it just doesn’t feel right to call an ant colony intelligent. Sure it can do whatever a self contained, decently intelligent animal would do, but where’s the feeling? Where’s the brain? For us there’s no such thing as just “intelligence”; some mystical force arbitrarily haunting its way through certain processes. There’s got to be a subject that intelligence can fill up like a glass of water. Yet most of us have no problem with the many smarty pants antics of artificial intelligence. The idea that we can set up a certain number of programmatic pieces which will go on to solve their own problems seems only natural to us. These things might be complicated, but they are finally just elaborate mouse trap contraptions with a human being at the bottom of them. Of course intelligence can make other intelligence. It may have taken us a while to get smart enough to animate that process, but there’s a clean line we can draw back to a big juicy brain that makes us alright with leaving the intelligence label on what these programs do. 

     But artificial intelligence and ant colonies have a lot more in common than human intelligence and artificial intelligence do. For one thing artificial intelligence is basically the result of a human swarm. From the concept of programming, to the chips that run the programs, to the raw materials that build those chips, artificial intelligence is something that emerges from a busy body hive of human activity. For another, artificial intelligence is as widely spread and idiosyncratic in the way it works as any ant colony. In the same way that ants are perpetually scaling the walls and combing the dirt on missions for the hive, artificial intelligence is for the most part combing the landscape of human activity, pinching up our data as it crawls across our digital trash in endless waves of bots with simple, atanae like tools to know what they’re picking up. 

Who’s thinking about who here? 

     Most ants and most artificial intelligence are incredibly simple little programmatic devices. What makes us recognize both of them as the independent authors of insightful solutions, is the larger continuity of meaning that rises out of their many individually simple actions. Really, the same thing happens in any singularly intelligent animal. A bunch of cells do a bunch of individually simple jobs and rising out of the connections between their purposes is the thing you see when you look in the mirror. People can mess with your brain like we could mess with the queen of an ant colony, but the queen and the colony are not the same thing. It’s the shared purpose running through them that makes the intelligence what it is.

      Seeing this continuity of purpose artificially used to recreate intelligence in a digital form, forces us to pull back a little with the intelligence story we normally push at each other. Intelligence isn’t something that’s simply poured into organisms. It’s a process that arises when a whole collection of purposes presses itself into a landscape. There’s a continuity of purposes that solves life into systems through this scattered pressing of tiny operators, and those systems are as intelligent as they are coordinated. It’s only this pulled back perspective that can successfully describe the many faces of intelligence we see as it peeks out at us from colonies of ants, to strings of digital code, to our own bodies. 

     But using this as the definition of intelligence feels a little like drawing a mustache on the pope. How can we take this thing that’s supposed to make us a uniquely valuable creature and not only take it away from people, but divorce it from biology itself? But shifting the way we think about what intelligence is doesn’t do anything irreverent to the holiness of the human experience. In fact it’s just the opposite. Once we understand how our own intelligence appears, intelligence stops being some cracker jack prize that people carry in the center of their brains and starts being something that simply ripples into existence once the right kind of path is stretched out for it. By understanding how to draw that path we gain a new command of our own most valuable dimensions and renew what it means for life to matter when we open our eyes. 

And they can change the directions they fall in as the environment changes around them. 

     Rethinking intelligence this way also gives us the opportunity to gain new insights on some very old questions. For instance, when we ask questions like: “What is knowledge?” we’re typically bringing a checklist to this or that idea. We look at something someone says they know and ask how logically sound that idea is or how verifiable it is. That’s because normally when we talk about knowledge, we’re trying to distinguish it from things like guessing or faith, which is a perfectly useful thing to do. But when we look at ant colonies or AI we’re seeing another face of knowledge; knowing as a phenomenon, the note that’s struck across what we are when we solve some piece of life out of the planet. 

     Intelligence is a kind of clicking in, a turning on. It’s what shows us a colony and not just a pile of ants. Once any collection of actions develops its own internal set of rules that recreate that collection, the shape carved out by those rules has intelligence inside of it. We tend to focus entirely on what intelligence can do, measuring it with tests or mapping its activity with medical equipment, and while there’s nothing wrong with that line of thought, it misses this moment of inception when an inside first ripples through a shared set of purposes. 

     From this perspective, knowledge is what happens when we add a gear to the internal process that originates with intelligence. Intelligence is a collection of impressions. That’s what ants are doing; scanning for ant impressive things. It’s what AI is sniffing around for and it’s how our own perceptions come into the world, by pressing some part of us into the world that ripples an impression of that world across our shared collection of cells. But knowledge is an act of managing our impressions. We have an idea, and then we know or don’t know it. We say; ‘this idea is a dream, this one is a fact, I’ll put this one over here, compare it to these other ones.’ The more we do it, the more we create the matrix of knowing that we call our individual selves. 

Knowing is bigger than brains

It’s the pattern not the place we find it. 

    One person can’t invent a language really. And even more than that, it seems like a panel of people tasked with the job of inventing a language would probably have a pretty tricky time with it. Yet groups of people do it organically, or even more to the point, groups of people behaving organically can’t actually stop themselves from doing it, without some intervening national system that they have to conform into. Without schools and laws to hold an official language in place, languages splinter off into countless regional tongues, not because any committee decides that they will, but because language is a way we have of knowing each other. We hang words on shared moments like little dibs signs that we get to come back for. We don’t know what people who speak other languages are saying, it’s because you had to be there when the inside first rippled through the joke. 

     A language is something we can know or not know. But it’s not just waiting out there in the world like some rubber ball that we can grab and start knowing about. Language is created in us through shared experiences. We can think in it, yet it’s only interactively true. So why should our panels and our sit down egg heads be so much worse at inventing languages than groups of people who aren’t even trying? If knowledge is an added layer of intelligence, then why is there this intellectual task that intelligence seems to accomplish without our knowledge? 

      An intelligent creature can’t help being intelligent. It wakes up, it sees, it understands. But knowledge has to be built. We have to inspect things, to meditate on the impressions they leave. Intelligence is something we can’t lose, but knowledge is a much more active process. This seems to mark out a split between the two, as though something altogether foreign enters the picture when we start knowing about things. But knowledge is a clear continuation of the same process that ripples intelligence through our insides. It’s the direction our colonies are moving in. If one person knows and another doesn’t, the person with the knowledge picks the direction the two of them will go in. When we argue about who knows what, we’re in a little tug of war for which ways our colonies are going to move. Knowledge is an exchange of impressions, a way of managing their storage that decides which ones will go where. It’s a scaling up of the process we’ve recreated in artificial intelligence, the same process that our senses use to sniff our first thoughts out of the world. 

I’ve just gotta find a gayer shoe store. But where? 

    Individual intelligence is a synthesis of purpose in the form of our cells growing together. Swarm intelligence is a synthesis of purpose created by identical things acting together. And cultural intelligence, the kind we find in shared linguistic meanings, is a synthesis of individual purposes; basically swarm intelligence phase two. In each case the intelligence can be more fully synthesized, like when that synthesis folds over on itself to become knowledge in individuals. But it’s never some appendage inherited by creatures. It’s what happens when a certain process hits a rolling boil. 

Think swarm 

A big job is actually a bunch of little jobs. 

    We talk about our countries like they’re somewhere else. This guy’s ideas will hurt the country. That guy’s ideas are un-American. From an individual perspective, intelligence is the passive quality of waking life and knowledge is an active thing that grows out of applying intelligence. But from the perspective of our groups, the knowledge we sniff out for them is the passive receptacle of their collected intelligence. Our countries passively exist between us. A law is everywhere, our  history is inherited. True success is defined by the moment our active contributions leave the control of our personal momentum and become social facts that passively define us. Actors want to become famous, where the knowledge of them stops being something they actively produce and enters a kind of cultural entropy that reproduces itself. The same is true for scientific findings, for pivotal supreme court decisions. Crossing over from the active effort of individual knowing to the passive intelligence that exists between us as an inherited meaning is a shared project of all colony members. We produce, spread and digest each other’s truths, until that spreading around creates the character of what we are. 

“You’ve never seen the Godfather?” = The first thing everyone says to people who say they’ve never seen the Godfather.  

     This seems to fall in line very neatly with the scaling up process we’ve looked at so far. We’re basically gigantic ants with a bunch of turbo charged features, things like souls and morals, that let us hammer out gigantic colonies for ourselves with corresponding devices, things like constitutions and laws. We use processes like science and math in specialized burrows where we tinker and exchange with each other before bringing out directions that can spread through the hive. There appears to be a very direct order of operations here. The things we know collectively are actively taught, but they’re passively known. Our history, languages and the basic premises of things like math and chemistry are not the property of any active creature. They were produced at some point, but nobody really sat down and spelled them all the way out in a notebook somewhere. They passively arose out of our collected activity. 

     But fairly recently we’ve figured out how to use this collected process of figuring things out in an active way. 

Things are getting really touchy feely. 

     A lot of the coordination that needs to ripple across the mega hives we call countries has created some very expensive problems. Electricity gets to us through wires, but we built most of our communities before there was electricity and the task of living together is a messy business. As a result power stations have to be built and placed in very asymmetrical ways. Figuring out the shortest possible distance that wires can stretch across this scattered hive of activity can be a million dollar question. But it turns out our usual tools of math and science are not all that great at figuring out the solutions to these problems by themselves. There are a huge number of possible paths between electrical towers and calculating which one is the shortest is like punching in every lottery ticket number until you win. But people with dollars to spend have figured out that they can use digital swarms of ants to solve these kinds of problems

    Ants follow one another around because when one of them finds something worth walking to, they secrete a pheromone. The stronger the pheremorne the more ants are attracted to that place. But the longer an ant has to walk back to other ants, the weaker their pheromone trail becomes, like a dipped quill that’s run out of ink. The shorter the distance an ant has to walk, the more concentrated their trails are, and the easier other ants can find and follow that trail. So a giant swarm of ants who all follow this rule will organically crowd across the stronger trails, secreting more pheromones as they do, until the swarm organizes the shortest possible paths for itself across a given area. By setting up power stations on a computer program and throwing a swarm of digital ants at them, power companies stop trying to win the lottery with one at a time calculations and start knowing how to save themselves millions of dollars in wire costs. 

Methods in the madness. 

      A pretty critical reversal in our intelligence order of operations is happening here. In a lot of ways things like math and science are ways of unscrambling our swarm like behaviors. Our organic ways of organizing are not all that precise and we have to chop apart our first impressions back at the hive with scientific analysis and hard calculation to get those sweet knowledge goodies out of them. Yet the power lines problem is one that has to solve its way through our swarm-like patterns, rather than simply unscramble them. As people who were only trying to make sense out of a natural world, we had to develop the unscrambling tools that bring hard knowledge back from organic processes. But now some of our biggest jobs involve making sense out of our own organic ways of life, and that means finding a way to use scrambling precisely. 

     Swarm intelligence programs like the ones that solve the powerlines problem have math in them; the digital ants have algorithmic rules they use to sniff and secrete their way through the problem. But the process that actually solves for the cheapest amount of wire isn’t a mathematical one. So what’s happening as we scale up the interconnected structures of our colonies is that we’re gaining an active control over what were formerly only passive forms of knowledge. Intelligence and knowledge exist on a continuum, but the point of departure when the passive rippling of intelligence crests over into the breaking wave we experience as conscious life is the result of an internal surging. A ripple comes out of an ocean, but a wave breaks over itself. As we meditate on our own organic processes we become aware in new ways. 

      What I’d really like to suggest here is that there are more ways to know things than math and science make available by themselves. We are still very much coming into our own as a species, a process underscored by the fact that swarm intelligence solutions to the power line problem and others like it bear a striking resemblance to the neurological patterns that ripple intelligence across the human brain. We tend to forget that math and science are ways of arriving at the truth, but they’re not the truth itself, which doesn’t particularly care what path we took to get to it. As we crowd our way across the planet and stand taller on the passing years of history, we’re going to need to project solutions through a multibillion person mess. To do that we’re going to have to find ways to scramble smartly and that means opening up new paths to the truth. Start by opening your own mind. 

Everything counts in large amounts. 

Leave a comment